L2000 problem

[Avinash_1]

One of our L-class systems panic'd two times recently. I have attached latest PIM dump (ts99) file. Just wondering any hardware experts out there would able to spot any problem.

Thanks.

[jung-duk park]

Hi, there~

I put this ts99 file into the HPMC Decoder for rp54XX and got the result as follow.

Please refers to this.

Looks like this is a 2 CPUs system
All timestamps are valid and within each others ..
There is at least one chassis code different from 6322. Let s look at it
The CPU 0 has a chassis code different from 6322.
Bit 0 is set in CPU status1. This indicates an HPMC condition
HPMC can be Runway timeout, Runway parity, Unexpected error
broad error, directed error or Data cache parity .
Bit 8 is set : This indicates a Data miss
Most likely the system was trying to access the memory. Let's look at the responder address in the CPU registers :
Yes indeed. The system responder address is 0xfffffffffed00000 which maps to the ASTRO chip (memory main controller).
Without any other data here the main suspect is the SYSTEM BOARD where the ASTRO is located .


http://hphwec07.cup.hp.com/cgi-bin/decode_n4000_main.pl

[jung-duk park]

Hi, there~

I put this ts99 file into the HPMC Decoder for rp54XX and got the result as follow.

Please refers to this.

Looks like this is a 2 CPUs system
All timestamps are valid and within each others ..
There is at least one chassis code different from 6322. Let s look at it
The CPU 0 has a chassis code different from 6322.
Bit 0 is set in CPU status1. This indicates an HPMC condition
HPMC can be Runway timeout, Runway parity, Unexpected error
broad error, directed error or Data cache parity .
Bit 8 is set : This indicates a Data miss
Most likely the system was trying to access the memory. Let's look at the responder address in the CPU registers :
Yes indeed. The system responder address is 0xfffffffffed00000 which maps to the ASTRO chip (memory main controller).
Without any other data here the main suspect is the SYSTEM BOARD where the ASTRO is located .


http://hphwec07.cup.hp.com/cgi-bin/decode_n4000_main.pl


Bye~

[Avinash_1]

Thanks very much for your time. I could not access the URL you gave. Any specific reason?

Regards

[Patrick Wessel]

The decoder Jung-Duk refered to is not in the public internet but in the hp intranet.

[Stefan Saliba]

I wonder how he has acess to intranet if he is not HP.

[Avinash_1]

Patrick & Jung-Duk,

Is there any public version of this decoder available anywhere? Do you know any HP external site (internet) has this tool available to access. It looks like a useful tool to read PIM dump on HP servers.

Regards,
Anil

[Patrick Wessel]

Anil,
I'm not aware that there are any plans to link the decoder to the public internet.